The speed of modern financial transactions has created a new battleground in the fight against fraud. While artificial intelligence is now a central tool in detecting suspicious activity, Visa is warning that relying on AI alone is no longer sufficient to combat increasingly sophisticated scams. Criminals are leveraging the same technologies as financial institutions, and are exploiting the speed of modern payment rails – and, crucially, human psychology – to quickly move funds and evade detection.
The shift requires a more nuanced approach, one that combines the power of AI with human expertise and coordinated intelligence sharing, according to Aman Cheema, Visa’s vice president and head of Global Professional Services, Risk and Security Intelligence Solutions. Cheema recently discussed these evolving threats in an interview as part of the “Visa Protect” series, highlighting the demand for a proactive and adaptive defense against fraud that exploits both technology and human vulnerabilities.
The core challenge, Cheema explained, is that fraudsters are no longer simply trying to breach technical defenses. They are increasingly focused on manipulating individuals into willingly authorizing fraudulent transactions. This “social engineering” tactic, amplified by the capabilities of generative AI, makes scams harder to detect because they often appear legitimate to automated systems. The immediacy of faster payment systems further complicates matters, allowing criminals to quickly drain accounts before intervention is possible.
“The scammers can do whatever is in the armory to get the scam [successfully completed], and the money’s taken instantly and it’s gone,” Cheema said. “To get that money back proves to be remarkably difficult.”
AI: A Necessary, But Not Sufficient, Defense
Artificial intelligence has turn into a cornerstone of modern fraud prevention, capable of analyzing vast amounts of transaction data to identify anomalies and potential threats. However, its widespread adoption has also leveled the playing field, giving criminals access to similar tools. Which means that simply deploying AI is no longer a differentiator. Financial institutions must now layer additional defenses on top of AI-powered systems to stay ahead.
The key, Cheema argues, is to pair AI’s analytical capabilities with the judgment of experienced risk experts. These experts can interpret emerging patterns, connect seemingly disparate transactions, and respond quickly to coordinated attacks – tasks that automated systems often struggle with. This human element is particularly crucial in identifying scams that rely on social engineering, as these often bypass traditional fraud filters.
The Rise of Speed and Social Engineering
Modern scams are characterized by both their speed and their reliance on manipulating human behavior. Faster payment rails, while offering convenience, provide criminals with a critical advantage: the ability to move funds almost instantly, making recovery extremely difficult. According to the Federal Trade Commission, reports of fraud involving scams are steadily increasing, with losses totaling billions of dollars annually. The FTC reported over 3.3 million fraud reports in 2023, with a median loss of $500 per incident.
This speed is coupled with increasingly sophisticated social engineering tactics. Criminals are using generative AI to gather personal information, mimic trusted contacts, and craft highly persuasive communications designed to trick victims into authorizing payments. These scams often exploit emotional vulnerabilities, creating a sense of urgency or fear to bypass rational decision-making. Because these scams rely on authorized transactions, they often appear legitimate to automated monitoring systems.
Behavioral Analytics: Building a Customer Profile
To combat these evolving threats, financial institutions are turning to behavioral analytics. Visa’s Visa Protect, powered by Featurespace, analyzes transaction patterns to develop behavioral profiles for each customer. This allows the system to identify deviations from normal spending habits, flagging potentially fraudulent activity for review.
“We model your behavior on your payment transactions,” Cheema explained. “Systems analyze purchasing patterns to understand what a customer’s normal behavior looks like.” For example, a sudden large transaction from an unfamiliar location or device would trigger an alert. However, Cheema emphasized that the challenge isn’t simply collecting data, but identifying the *meaningful* signals within the noise.
The Power of Shared Intelligence and Human Oversight
Filtering out the noise requires a broader ecosystem approach, with shared intelligence among issuers, merchants, and payment providers. By sharing information about emerging fraud patterns, institutions can detect coordinated campaigns more quickly and respond before they spread. This collaborative approach is essential, as criminals often target specific customer profiles or regions in organized attacks.
“You see patterns in one part of the world which make you immediately think that if they’re attacking that profile, there’s a very high likelihood they’re going to be attacking the same persona profile in another part of the world,” Cheema said.
Cheema believes the most effective fraud prevention strategy combines “humans, applications, and data,” each strengthened by AI. Visa’s fraud specialists actively monitor for suspicious activity and can directly contact issuers, merchants, or acquirers to confirm potential threats and provide guidance. This proactive approach is crucial in mitigating the impact of increasingly sophisticated scams.
The fight against fraud is a continuous evolution, requiring constant adaptation and innovation. As criminals continue to refine their tactics, financial institutions must prioritize a layered defense that combines the power of AI with the critical thinking and expertise of human analysts. The next step in this ongoing battle will likely involve even greater collaboration and data sharing across the payments ecosystem, as well as continued investment in technologies that can detect and prevent these increasingly complex attacks.
Have thoughts on the evolving landscape of fraud prevention? Share your comments below, and let us grasp how these changes are impacting you.
